An investigation of periglacial slope stability in relation to soil properties based on physical modelling in the geotechnical centrifuge

Results are presented from eight scaled centrifuge modelling experiments designed to investigate mass movement processes on thawing ice-rich slopes. Four pairs of simple planar slope models were constructed, one in each pair being of sufficient gradient to promote slope failure during soil thaw and the second having a gradient below the threshold for instability. Four frost susceptible soils were used, three were normally consolidated and had different clay contents (2%, 12% and 20%) and the fourth comprised the 20% clay soil, but was over consolidated prior to model testing. Modelling protocols included freezing from the surface downwards under an open hydraulic system, and thawing from the surface downwards under an enhanced gravitational field within the geotechnical centrifuge, thereby utilising scaling laws to simulate correct prototype self weight stresses during thaw. Slopes below the stability threshold gradient were subjected to between 2 and 4 cycles of freezing and thawing, simulating annual cycles. Those above the stability threshold were subjected to only one cycle, since they failed during the first thaw phase. Thermal conditions, pore water pressures, surface movements, and profiles of displacement are reported. Measured pore pressures are used in slope stability analyses based on a simple planar infinite slope model. Profiles of solifluction shear strain and mechanisms of slope failure are both shown to be sensitive to small changes in soil properties, particularly clay content and stress history. In all cases, pore pressures rose rapidly immediately following thaw, remained below the threshold for failure in low gradient models, but exceeding the threshold to trigger landslides on steeper slopes. Upward seepage of melt water away from the thaw front contributed to loss of shear strength. Mechanisms of slope failure differed between test soils, ranging from mudflow in non-cohesive silt to active layer detachment sliding in over consolidated silt–clay. During solifluction, shear strain was greatest at the surface in non-cohesive silt and decreased rapidly with depth, but in test soils containing clay, the zone of maximum shear strain was located lower in the displacement profiles